Locus-specific sibling relative risk is often estimated using affected-sib-pair lod score analysis of affected sibships and may be used to decide whether to continue or discontinue the search for additional susceptibility genes. We show that relative risk estimates obtained using affected-sib-pair data are asymptotically unbiased only when each pair is given a weight inversely proportional to the sibship ascertainment probability. Simulations show that the extent of bias of relative risks estimated using the incorrect ascertainment assumption is small for dominant models, but large for single-locus recessive models and some two-locus heterogeneity models. Since in practice the ascertainment scheme is usually unknown, we investigated methods for jointly estimating ascertainment and relative risks from affected-sibship data. Given a sufficient sample size, a reasonable estimate of relative risk may always be obtained using only affected pairs from sibships with two affected and no unaffected siblings. This estimate, which has a large variance, may then be used in a three-stage procedure to consistently estimate both the ascertainment probabilities and the relative risks with greater precision. We also propose correction factors to eliminate small-sample bias and investigate the bias due to error in the estimate of disease locus location.